The Best Pricing Model for non-FEM Structural Software - KootWare 5

[KootK]

The Mission

While I'm still above grade, I intend to create a suite of pay for play, online structural engineering tools (KootWare). And I feel that a big part of making this questionable venture a success -- or at least improving the odds of a contained failure -- will be arriving at a good pricing model. Frankly, this is something that I feel that other developers have done poorly, to their detriment. As such, I'd like to solicit feedback from the hive with respect to the pricing models that I'll propose below and any possibilities for improvement.

The Basics of What You Need to Know About the Offering

1) 100% online offering. No option for a local, perpetual license version.

2) The goal here is not to get rich. The goal is to extract enough income from this that I can justify pouring a lot of effort into a project that I expect to enjoy a great deal.

3) Spit balling, if I could create enough value that I could convince 1000 SE's to part with $5/month, that would be enough. Or any other combination of numbers that gets to the same place. How many software using structural engineers do we think exist in North America anyhow? Sixteen? Eighty thousand? I really don't know.

4) Think something along the lines of TEDDS, ENERCALC, or Jabacus on steroids. I do have ideas for, in my opinion, greatly improving upon these offerings. I'd like that to be a separate conversation however. For now, make a leap of faith and just assume that it will be awesome.

5) I intend to attach some manner of structural only, online forum to the offering. While it would be a free-form space for conversation, as Eng-Tips is, it's ostensible purpose would be to provide a place for me to provide responsive help to anybody designing stuff utilizing the software. Thus making the whole thing even more fun for me. This would be offered in addition to the usual help guide and verification manuals etc <-- edit added per skeletron's comments.

Some Obvservations that I Have Regarding the Pricing Models of Others

6) For software of this type, I feel that a monthly subscription pricing scheme would not be well received. As a small outfit my self, I loathe taking on any additional "monthlys", no matter how great the ROI seems to be. I'm always afraid that I'll use it twice and forget to cancel. I doubt that I'm the only one who feels this way.

7) I also don't think that a straight "pay per use" model is the way to go either. Design is an iterative process and software licensing needs to reflect that. Sadly, I don't just design a shear wall once. I probably design it half a dozen times before all is said and done. And I can't be losing my shirt on pay per use while going through that process.

8) One has to assume that anything that can be abused, will be abused. This will prevent me from being quite as customer friendly as I would otherwise wish to be. My own IP halo gets a little dirty from time to time so no judgement here.

Pricing Model A

This is my favorite of the two and would appeal to me as a customer. Keep in mind than none of the particular values are set in any way. It's really more about the structure at this point. That said, if anybody has thoughts on what the numbers ought to be, I'd welcome that too. I figure I'll adjust as use data starts to pile up but I'll still have to start somewhere.

1) Create an account at KootWare International and add a credit card, paypal etc.

2) Buy yourself some quantity KootWare credits. $10. $100. Whatever. Little gold doubloons in your digital purse.

3) To access the retaining wall tool for use, you pay $5. After the first run, you have the lesser of 20 additional runs or 60 days to keep using the tool on the original $5. One "run" would represent one execution of a full design with detailed output. <-- added per skeletron's comments.

4) If you want to share your account login and credits with somebody else, that's your prerogative. Share it with your coworker, a school chum in Brisbane, your aunt... retaining walls for everybody on that original $5. But, no matter who's using, it taps out after 20 runs or 60 days.

Pricing Model B

1) Create an account at KootWare International and add a credit card, paypal etc.

2) Buy yourself some quantity KootWare credits. $10. $100. Whatever.

3) You can use any tool your like, for free, but you can't get a detailed printout for your calcs until some money has changed hands. The software would tell you the basics of what passed and what failed and would allow you to save your file to the system for future retrieval. I kind of like this in that it would allow one to essentially do their preliminary design work for free. I could allow folks to printout their inputs in case they were worried about my going bankrupt before they get to IFC.

4) When you've got all your design settled and ready for final calc documentation, it's $5 per print. The trouble with this is, I couldn't let the user see the detailed printout ahead of paying for it. Otherwise, I'll wind up with a bunch of folks just doing screen capture etc.

 

[bootlegend]

Pricing Model D


My Thinking

This is probably the best pricing scheme I've read yet. I was in the camp preferring a monthly subscription, but if you can offer the month of unlimited use for a fair price I think you'll get some of us on board for the a la carte system. I've demoed software before that had certain values locked down. I would rather pay for the trial period and get full capabilities to see if the software is right for me. Sounds like your pricing model will allow that.

I agree with point C as well. If the price is fair, the software is good, and it's easy to pay, etc I think most people are going to just pay for it when they need it rather than put forth the effort to game the system.

 

[KootK]

This is probably the best pricing scheme I've read yet.

Thanks so much for your input. Just what I need.

I agree with point C as well. If the price is fair, the software is good, and it's easy to pay, etc I think most people are going to just pay for it when they need it rather than put forth the effort to game the system.

Excellent. It's a relief to hear that I might not be nuts for considering this.

In reflecting upon Pricing Model D, maybe I can provide guest access as winelandv suggested. The gross abuse model that I outlined in my last post is probably a pretty unlikely outcome as well. Additionally, I'm sure that there's some way that I can roughly track the locations at which the software is being used. If things appear to be on the up and up, I'd leave the guest access feature in place as I think that would be great other than the abuse potential. If one account appears to be bouncing around from Pittsburgh to Cairo to Prague, I'll lock it down.

HELP! I'd like your help with a thread that I was forced to move to the business issues section where it will surely be seen by next to nobody that matters to me:

http://www.eng-tips.com/viewthread.cfm?qid=456235

 

[Celt83]

I like model D.

To address the guest account thing I think it'd be better to limit it to an admin account with purchase power and company user accounts tied to email address domains so if company B buys Kootware Retain Master then the admin registers the other users on Kootware as user account with a technical limit that any user account must have the same email address domain as the admin account, admin@companyB.com and user@companyB.com.

edit:
potential side benefit of doing it this way is you can get all kinds of metrics on the use and can share the per company metrics with the admin accounts.

Open Source Structural Applications:

https://github.com/buddyd16/Structural-Engineering

 

[Agent666]

Yeah having someone administer sub-accounts is good, or lock access to company IP's, or some other method. Otherwise you get users who leave the company have free use if they still have a valid login....

 

[KootK]

Master then the admin registers the other users on Kootware as user account with a technical limit that any user account must have the same email address domain as the admin account, admin@companyB.com and user@companyB.com.

That's brilliant, thank you.

HELP! I'd like your help with a thread that I was forced to move to the business issues section where it will surely be seen by next to nobody that matters to me:

http://www.eng-tips.com/viewthread.cfm?qid=456235

 

[KootK]

Yeah having someone administer sub-accounts is good, or lock access to company IP's, or some other method. Otherwise you get users who leave the company have free use if they still have a valid login....

Great point.

HELP! I'd like your help with a thread that I was forced to move to the business issues section where it will surely be seen by next to nobody that matters to me:

http://www.eng-tips.com/viewthread.cfm?qid=456235

 

[KootK]

After all that has been discussed, it's almost a bit sad that this might just boil down to:

Monthly, per module, no auto-renew.

Although, it is usually the case that the most elegent solution winds up also being simple.

I suppose that I shouldn't get too excited jut yet though. Only three nods for Model D so far. I assume that windlandv is more or less on board pending the guest access refinement.

HELP! I'd like your help with a thread that I was forced to move to the business issues section where it will surely be seen by next to nobody that matters to me:

http://www.eng-tips.com/viewthread.cfm?qid=456235

 

[WinelandV]

I mean, depending on the pricing, I might just subscribe for myself. I find myself open to most of the pricing models, just because I think this is a worthy endeavor.

 

[KootK]

I'm getting way ahead of myself with this but it's just...so...much...fun.

Now that the monthly and al a carte options have pretty much collapsed into a single thing, I feel that I can see in my head how purchasing would be structured in terms of the GUI. There would be three options:

1) Buy single tool access right from that tool.

2) Set up an auto renew thing as shown in the first clip below.

3) Set up a batch, one time purchase thing as shown in the second clip below.

HELP! I'd like your help with a thread that I was forced to move to the business issues section where it will surely be seen by next to nobody that matters to me:

http://www.eng-tips.com/viewthread.cfm?qid=456235

 

[Agent666]

I feel you may be colourblind and need to work on your colour scheme.... (I realise it's just a mock-up, but my eyes are almost bleeding here)

 

[KootK]

Well, I'm sorry for any damage to your eyes. Just a concept thrown together as quickly as I could.

HELP! I'd like your help with a thread that I was forced to move to the business issues section where it will surely be seen by next to nobody that matters to me:

http://www.eng-tips.com/viewthread.cfm?qid=456235

 

[WillisV]

If you spend as much time developing your business model and software as you do posting on eng-tips you will surely be filthy rich.

 

[BUGGAR]

I, too, am fascinated by your strut and tie idea. I have been playing with my own re-invention of the typical auto suspension geometry analysis combined with structural analysis of the parts. I think this is similar, or adaptable to your stm. Its doable, but my hang up is joining Autocad and excel.

 

[KootK]

.If you spend as much time developing your business model and software as you do posting on eng-tips you will surely be filthy rich.

That's the idea. Spend my time on things that I care about intensely and stop counting hours.

HELP! I'd like your help with a thread that I was forced to move to the business issues section where it will surely be seen by next to nobody that matters to me:

http://www.eng-tips.com/viewthread.cfm?qid=456235

 

[KootK]

I, too, am fascinated by your strut and tie idea.

Somewhat counterintuitively, I see the key to my STM thing being intentional imprecision. Something like this:

1) Import a DXF PDF of the thing to be modeled, the load points, and the reaction points. Trace it.

2) Run a generic elastic FEM on the model just to give the user an idea of where the load paths "want" to be.

3) Draw a free form, imprecise truss over the model to serve as the STM.

4) Draw in nodes free form and imprecise. Triangles etc that you stretch and move the vertices of.

5) As #3 and #4 happen, the STM forces, strut checks, and nodal checks update in real time.

6) Some kind of nifty metric is calculated that quantifies how closely STM load paths mimic FEM load paths.

It is my belief that most STM software attempts fall short in forcing the user to create precise models which, given the complexity involved at the nodes, takes forever and is tedious. With the method outlined above, I suspect that an engineer can quickly come up with a rough-ish model suitable for design purposes in very little time. And there, I hope, will lie the power of it.

HELP! I'd like your help with a thread that I was forced to move to the business issues section where it will surely be seen by next to nobody that matters to me:

http://www.eng-tips.com/viewthread.cfm?qid=456235

 

[WillisV]

KootK - Star for that response and your fire for creation - best of luck.

 

[Agent666]

I bet there is plenty of open source machine learning algorithms that can optimise that crude truss model as well, or suggest better alternative arrangements through a optimisation/generative design process. I mean programs like Galapagos or Dynamo/Refinery and others already exists for optimising and optioneering structural form based on constraints. Why not the same for Strut and tie! As its simply drilling down further looking inside a member to track the most efficient loadpath (in very simplified terms). So I perhaps naively think some kind of machine learning could be used with suitable constraints like max strut angles, ties parallel to member faces, automatically generating nodal zones, etc to generate an acceptable model.

A bit similar to some of the stuff you see on youtube where they use Autodesk Fusion 360 (I think) to optimise a given part for where the load wants to go to get rid of unnecessary material.

Making a S+T software that allows you to drag and drop nodes around in real time is something I feel the industry is actually missing, and I'd probably part with some money if I had a need from time to time. I'm surprised someone in the world hasn't thought of it before, we've got 101 analysis programs for frames and walls, but nothing really compelling in the S+T space. For me strut and tie is basically a hand calculation exercise for the most part, and choosing the 'right' model requires experience about how it is really going to work as you've noted.

Unfortunately tools that seem to come out of Universities (like CAST) that I mentioned, do suffer from the fact that the user interfaces can be very clunky. I suspect though you'd be charging more than $5 a pop for that given the steeper development time.

Just thinking outside the box, one other option that could be worth considering is the use of a kickstarter campaign or similar, get some people in on the ground floor to fund it. Offer them free use for a few years or something or discounted rate for a few hundred dollars or something per tool/set of tools. They fully or partially fund the initial development which you get done by people in the know how. Then once developed, new subscribers follow your chosen pricing model, if you can provide the value for the initial investment then it might be an attractive option for some companies/individuals. Maybe once you have the initial tools, then people can see the benefit of the approach and/or quality they might be getting. Could be a bigger ask to get people to invest with nothing in the pipeline though at the beginning.

 

[glass99]

Agent666 - we do that kind of optimization mostly in a Strand7 API. You can do some fun things with beam elements with custom scripts. The off the shelf optimization software is overly geared to mechanical engineering applications.

The Rhino environment is perhaps a good one for low budget software development too bc you start with a bunch of CAD tools, and a platform to sell plugins. What % of structural engineers have Rhino? I know we use it a lot.

Agree that KootWare should generate the struts and ties for you.

 

[KootK]

If you create a good tool, maybe you really kick-start the STM analysis movement.

That's an interesting point and you may well be right. Chicken and egg.

Agree that KootWare should generate the struts and ties for you.

I actually disagree. I feel that to go in the direction of genetic optimization algorithms etc is to slip back into making the thing an academic science experiment rather than a practitioner's production tool. Additionally, virtually every STM that I've done for real world work ended up being configured to suit what I knew would be a practical reinforcing layout, even if that wasn't 100% optimal from a material or load path efficiency perspective.

A routine to auto suggest the struts and ties could surely be added to the base feature set. But I would consider that a "for sport" addition rather than a mission critical one. And, really, the elastic FEM that I proposed running pre-STM-build should cover a fair bit this functionality. Any designer unable to translate the FEM stresses into a reasonable model probably shouldn't be using the tool anyhow.

HELP! I'd like your help with a thread that I was forced to move to the business issues section where it will surely be seen by next to nobody that matters to me:

http://www.eng-tips.com/viewthread.cfm?qid=456235

 

[jayrod12]

Any designer unable to translate the FEM stresses into a reasonable model probably shouldn't be using the tool anyhow.

That could be said for almost every tool ever created.